U.S. patent application number 10/896895 was filed with the patent office on 2005-02-24 for circuit-pattern inspection apparatus.
Invention is credited to Hayakawa, Kouichi, Hiroi, Takashi, Nara, Yasuhiko, Nojiri, Masaaki.
Application Number | 20050043903 10/896895 |
Document ID | / |
Family ID | 34189765 |
Filed Date | 2005-02-24 |
United States Patent
Application |
20050043903 |
Kind Code |
A1 |
Nara, Yasuhiko ; et
al. |
February 24, 2005 |
Circuit-pattern inspection apparatus
Abstract
The disclosed subject matter is related to a circuit pattern
inspection apparatus for detecting a gradual changing of defect
expanding over a large area of the semiconductor wafer. In order to
detect a gradual changing of a defect related condition expanding
over a large area of the semiconductor wafer, comparison is made
between dies on a wafer that are separated from each other by a
distance of at least one die width. For example, when a value
according to a difference between such dies exceeds a
pre-determined value, an existence of the gradual changing can be
confirmed.
Inventors: |
Nara, Yasuhiko;
(Hitachinaka, JP) ; Nojiri, Masaaki; (Hitachinaka,
JP) ; Hayakawa, Kouichi; (Hitachinaka, JP) ;
Hiroi, Takashi; (Yokohama, JP) |
Correspondence
Address: |
MCDERMOTT WILL & EMERY LLP
600 13th Street, N.W.
Washington
DC
20005-3096
US
|
Family ID: |
34189765 |
Appl. No.: |
10/896895 |
Filed: |
July 23, 2004 |
Current U.S.
Class: |
702/35 |
Current CPC
Class: |
G06T 7/001 20130101;
G01N 2021/95615 20130101; G06T 2207/30148 20130101 |
Class at
Publication: |
702/035 |
International
Class: |
G06F 019/00 |
Foreign Application Data
Date |
Code |
Application Number |
Jul 25, 2003 |
JP |
2003-201458 |
Claims
What is claimed is:
1. A circuit-pattern inspection apparatus comprising: a beam
source, a detector for detecting signals obtained from beam
irradiation of a part of a specimen, a memory for memorizing the
signals as one or more images, an operation unit for arbitrarily
selecting a die to be inspected and a reference die which is not
adjacent to any side of the die to be inspected, and a processing
unit for comparing an image memorized in the memory of the die to
be inspected with a reference image memorized in the memory of the
reference die, and for discriminating defects on the die to be
inspected based on a result of the comparison.
2. The circuit-pattern inspection apparatus according to claim 1,
wherein the operation unit sets the reference die separated from
the die to be inspected by at least one intervening die.
3. The circuit-pattern inspection apparatus according to claim 1,
wherein the operation unit for further selecting a die to be
inspected and a reference die which is adjacent to any side of die
to be inspected.
4. The circuit-pattern inspection apparatus according to claim 1,
wherein the beam source is a charged particle beam source, and the
detector is an electron detector for detecting electrons emitted
from the specimen.
5. The circuit-pattern inspection apparatus according to claim 1,
wherein the beam source is a light source, and the detector is a
light detector for detecting a reflection beam reflected from the
specimen.
6. The circuit-pattern inspection apparatus according to claim 1,
further comprising a stage for moving the specimen, the stage
controlled by the processing unit.
7. The circuit-pattern inspection apparatus according to claim 6,
wherein the processing unit controls the stage according to a
selecting of the die to be inspected and the reference die by the
operation unit to move the die to be inspected and the reference
die to an irradiation point of the beam.
8. The circuit-pattern inspection apparatus according to claim 1,
wherein the processing unit calculates a correlative value between
the image of the die to be inspected and the reference image.
9. The circuit-pattern inspection apparatus according to claim 1,
wherein the operation unit enables changing of an inspection order
of the dies to be inspected after the setting the inspection
order.
10. The circuit-pattern inspection apparatus according to claim 1,
wherein the operation unit can add or delete a die to be inspected
after the setting the dies to be inspected.
11. The circuit-pattern inspection apparatus according to claim 1,
wherein the operation unit enables selection of a plurality of the
other dies against the die to be inspected.
12. A pattern inspection method for detecting a defect on a
semiconductor device comprising: irradiating a beam to a die to be
inspected on the semiconductor wafer and a reference die which is
remote from the die to be inspected on the semiconductor wafer,
forming a first image of the die to be inspected and a second image
of the reference die based on signals generated from an irradiation
point of the beam, comparing the first image and the second image,
detecting a difference between the first image and the second
image, and determining whether the there is defect between the die
to be inspected and the reference die based on the deference
detected.
13. A circuit-pattern inspection apparatus comprising: a beam
source, a detector for detecting signals obtained from beam
irradiation of a part of a specimen, a memory for memorizing the
signals as an image, an operation unit for arbitrarily selecting an
order of dies scanned in sequence so that a die to be inspected and
a separate reference die are consecutive in the sequence, and
processing unit for comparing an image memorized in the memory of
the die to be inspected with a reference image memorized in the
memory of the reference die in accordance with the order set, and
for discriminating defects on the die to be inspected base.
14 The circuit-pattern inspection apparatus according to claim 13,
wherein the operation unit for arbitrarily selecting a die to be
inspected and the reference die which is not adjacent to any side
of the die to be inspected. a computer program product for
controlling a circuit-pattern inspection apparatus comprising a
beam source, a detector for detecting signals obtained from beam
irradiation of a part of a specimen, a memory for memorizing the
signals as an image,
15 A program product for controlling a circuit-pattern inspection
apparatus comprising a beam source, a detector for detecting
signals obtained from beam irradiation of a part of a specimen, a
memory for memorizing the signals as an image, and an operation
unit for selecting a die to be inspected and a reference die
compared with the die to be inspected on a semiconductor wafer, the
product comprising: executable code embodied in a readable medium,
execution code causing the apparatus to perform a sequence of steps
comprising: (a) irradiating a beam to a die to be inspected and a
reference die which is remote from the die to be inspected by
setting the die to be inspected, the reference die, and an order of
dies scanned in sequence so that a die to be inspected and a
separate reference die are consecutive in the sequence with the
operation unit, (b) forming a first image of the die to be
inspected and a second image of the reference die based on signals
generated from an irradiation point of the beam, (c) comparing the
first image and the second image, and (d) detecting a difference
between the first image and the second image.
Description
TECHNICAL FIELD
[0001] The present subject matter relates to techniques and
equipment for circuit-pattern inspection, used in manufacturing of
a semiconductor devices and photomasks.
BACKGROUND
[0002] In the manufacturing process of a circuit pattern formed on
the semiconductor device and the photomasks, it is difficult to
avoid the occurrence of a defects such as a production failure and
the adhesion of foreign materials. For this reason, the following
inspection apparatuses are put to practical use: an optical pattern
inspection apparatus that irradiates a target to be inspected with
white light, and that then compares between circuit patterns of the
same kind of a plurality of LSIs by use of an optical image to
extract a defect; and a SEM pattern inspection apparatus that
applies the scanning electron microscope technology, and that
thereby compares circuit patterns by use of an electron-beam image,
the resolution of which is higher than that of the optical image,
to extract a defect.
[0003] Because the inspection speed of this SEM pattern inspection
apparatus is slower than that of the optical pattern inspection
apparatus, the SEM pattern inspection apparatus is so devised that
instead of specifying a semiconductor chip (that is to say, all
dies) on a semiconductor wafer, a die to be inspected can be
specified (for example, refers to Japanese Patent Laid-Open No.
2000-162143 (FIGS. 13, 14)).
[0004] Japanese Patent Laid-Open No. 2000-162143 discloses that an
operator can specify dies as an area of inspection. The image
comparison checking is a method in which adjacent dies are
inspected and compared in sequence. The comparing between the
adjacent dies makes it possible to identify a defect. A pattern
inspection method of the prior art has performed by comparing
between adjacent dies, in a whole semiconductor wafer or in a part
of the semiconductor wafer.
SUMMARY
[0005] Although the prior art technique can identify the defect
based on detecting a difference between adjacent dies, it can not
detect a defect changing gradually extending over a large area of
the semiconductor wafer. Although a total charge or difference
across a wafer may be substantial and result in a defect, the
incremental difference from one die to the next may be too small to
discriminate as a defect.
[0006] An object is to provide: a circuit pattern inspection
apparatus for detecting a gradual changing of defect expanding over
a large area of the semiconductor wafer.
[0007] In order to detect a gradual changing of defect expanding
over a large area of the semiconductor wafer, comparison is made
between dies on a wafer that are separate, e.g. by one or more
intervening dies or by a distance of at least on die width. And for
example, when a value according to a difference between the dies
exceeds a pre-determined value, an existence of the gradual
changing can be confirmed.
[0008] A circuit-pattern inspection apparatus comprises: a beam
system having a beam source, a detector for detecting signal
obtained from beam irradiation part on a specimen, a memory for
memorizing the signal as an image, and an operation unit for
setting a die to be inspected on a semiconductor wafer and a die to
be used as the reference die. The reference die is separated from
the die to be inspected on the semiconductor wafer. A processing
unit compares an image memorized in the memory of a die to be
inspected with a reference image memorized in the memory of the
selected reference die, and discriminates defects on the die to be
inspected based on a result of the comparison the image of the die
to be inspected and the reference image.
[0009] According to above subject matter, a defect changing
gradually extending over a large area of the semiconductor wafer
can be detected by selecting the reference die that is separated
from the die to be inspected as the reference die. The defect
changing gradually extending over a large area of the semiconductor
wafer can be detected, because even if the changing of distribution
between dies is gradual, it is emphasized by comparing between
separated dies.
[0010] Furthermore the other object is to provide a circuit pattern
inspection apparatus for detecting a defect which a technique
comparing the die to be inspected with the reference die along an
arrangement of dies as the prior art can not detect. Because the
comparison technique between dies compares the die to be inspected
with the reference die, if there is almost no difference between
the die to be inspected with the reference die, the die to be
inspected is recognized as a correct die. But the defect may be in
existence in case of comparing the die to be inspected with the
other reference die.
[0011] For example, in case of the reference die having a same
feature as the die to be inspected according to the defect, it can
not discriminate the feature as a defect. This fact means that
there are defects can not be discriminated according to what order
of the dies the comparison is performed and which die is set as the
reference die.
[0012] In order to detect the defect can not be detected by the
technique comparing the die to be inspected with the reference die
along the arrangement of dies as the prior art, the comparison
between the dies is performed by setting a die being separated from
a die to be inspected as a reference die and an order of dies being
compared.
[0013] A circuit-pattern inspection apparatus comprises: a beam
source, a detector for detecting signal obtained from beam
irradiation part on a specimen, a memory for memorizing the signal
as an image, and an operation unit for setting a die to be
inspected on a semiconductor wafer and a die to be used as the
reference die and an order of comparison between the die to be
inspected and the reference die. The reference die is separated
from the die to be inspected on the semiconductor wafer. A
processing unit compares an image memorized in the memory of a die
to be inspected with a reference image memorized in the memory of
the selected reference die according to the order set by the
operation unit, and discriminates defects on the die to be
inspected based on a result of the comparison the image of the die
to be inspected and the reference image.
[0014] The further detailed composition and effects of the present
subject matters are made obvious in the following description of
exemplary embodiments.
BRIEF DESCRIPTION OF THE DRAWINGS
[0015] The drawing figures depict one or more implementations in
accord with the present concepts, by way of example only, not by
way of limitations. In the figures, like reference numerals refer
to the same or similar elements.
[0016] FIG. 1 is a longitudinal sectional view schematically
illustrating a configuration of a circuit-pattern inspection
apparatus;
[0017] FIG. 2 is a diagram illustrating a screen configuration of
an operation unit;
[0018] FIG. 3 is a flowchart illustrating processing of a recipe
creation mode;
[0019] FIG. 4 is a diagram illustrating a screen used in a recipe
creation mode;
[0020] FIG. 5 is a diagram illustrating a screen for setting the
inspection order of dies; and
[0021] FIG. 6 is a diagram illustrating a screen used for fixed
settings of a reference die.
DETAILED DESCRIPTION
[0022] In the following detailed description, numerous specific
details are set forth by way of examples in order to provide a
thorough understanding of relevant teaching. However, it should be
apparent to those skilled in the art that the present teaching may
be practiced without such details. In other instance, well known
methods, procedures, components, and circuitry have been described
at a relatively high-level, without detail, in order to avoid
unnecessarily obscuring aspects of the present concepts.
[0023] Embodiments of a circuit-pattern inspection apparatus will
be described in detail with reference to drawings as below. In this
embodiment, targets to be inspected include a resist pattern, a
CONT system's opening pattern, a Fine pattern after etching
(diffusion system), and a Fine pattern after etching (wiring
system). A defect on these targets is detected by comparing die
images.
[0024] FIG. 1 is a longitudinal sectional view schematically
illustrating a configuration of a circuit-pattern inspection
apparatus. A circuit-pattern inspection apparatus 16 comprises an
inspection cabinet 17 that is vacuum pumped, and an extra cabinet
(not illustrated) used to transport to the inspection chamber 17 a
substrate to be inspected 24. This extra cabinet is devised to be
capable of being vacuum-pumped independently of the inspection
chamber 17. Incidentally, the reason why the circuit-pattern
inspection apparatus 16 is configured in such a manner is because
an electron beam is used. Therefore, in the case of a laser beam,
or the like, it is not necessary to adopt a vacuum-pump structure.
In addition, the circuit-pattern inspection apparatus 16 includes
not only the inspection chamber 17 and the extra cabinet, but also
a control unit 21 and an operation unit 20.
[0025] Broadly classifying, the inspection chamber 17 comprises an
electro-optic system 18, a secondary electron detector part 62, a
target chamber 23, and an optical microscope 19. The electro-optic
system 18 comprises an electron gun 25, an electron-beam emitting
electrode 26, a condensing lens 27, a blanking deflector 28, an
iris 29, a scanning deflector 30, an objective lens 31, a reflector
plate 32, and an ExB deflector 33. The electro-optic system 18
forms an irradiation means.
[0026] Applying the voltage to a point between the electron gun 25
and the electron-beam emitting electrode 26 causes the electron gun
25 to emit a primary electron beam 34. Acceleration of the primary
electron beam 34 is achieved by applying the high-voltage
electronegative potential to the electron gun 25. The primary
electron beam 34 travels in the direction of the substrate to be
inspected 24 by the energy equivalent to the potential, and is then
converged by the condensing lens 27. Further, the primary electron
beam 34 is thinly narrowed by the objective lens 31 before the
substrate to be inspected 24 is irradiated with the primary
electron beam 34. The substrate to be inspected 24 is a substrate
having a minute circuit pattern, or the like. The substrate to be
inspected 24 includes, for example, a semiconductor wafer, a chip
(or a liquid crystal), and a mask. It is to be noted that a
scanning signal generator 59 for generating a scanning signal and a
blanking signal is connected to both the blanking deflector 28 and
the scanning deflector 30, and that each of the condensing lens 27
and the objective lens 31 is provided with a lens power supply.
[0027] A correction control circuit 58 monitors the discrepancy in
position and height by use of signals from a position-monitor
length measuring machine 48 and from an inspected-substrate height
measuring machine 49, and then generates a correction signal from
its result. After that, the correction control circuit 58 transmits
the correction signal to an objective-lens power supply 70 and the
scanning signal generator 59 so that a correct position is always
irradiated with the primary electron beam 34.
[0028] An image of a surface of the substrate to be inspected 24 is
first obtained by irradiating the substrate to be inspected 24 with
the thinly narrowed primary electron beam 34 to generate a
secondary electron 71, and then by detecting the secondary
electrons 71 in synchronization with scanning of the primary
electron beam 34 and moves of an X stage 47 and of a Y stage
46.
[0029] As for the secondary electron detector part 62, the
secondary electron detector 35 is placed above the objective lens
31 in the inspection cabinet 17. An output signal of the secondary
electron detector 35 is amplified by a preamplifier 36 that is
placed outside the inspection cabinet 17. Then, an A/D converter 37
converts the output signal into digital data. The target chamber 23
comprises a target stage 45, the X stage 47, the Y stage 46, the
position-monitor length measuring machine 48, and the
inspected-substrate height measuring machine 49. In addition, with
the object of fine tuning the rotation of the substrate to be
inspected 24, a rotatable stage (not illustrated) may also be
placed on the X stage 47.
[0030] The optical microscope 19 comprises a light source 50, an
optical lens 51, and a CCD camera 52. The optical microscope 19 is
placed at a position that is in proximity to the electro-optic
system 18 inside the inspection cabinet 17, and that is separated
from the electro-optic system 18 to such an extent that the optical
microscope 19 and the electro-optic system 18 do not exert an
influence on each other. The distance between the electro-optic
system 18 and the optical microscope 19 is made known so as to
ensure the compatibility of coordinates on the substrate to be
inspected 24 between the electro-optic system 18 and the optical
microscope 19. The CCD camera 52 detects reflected beam from the
substrate to be inspected 24. An image can be formed by output data
from the CCD camera 52.
[0031] The X stage 47 or the Y stage 46 permits a substrate to be
inspected 24 to reciprocate a known distance between the
electro-optic system 18 and the optical microscope 19. Moreover, it
may also be so devised that an angle at which the substrate to be
inspected 24 is irradiated with the primary electron beam 34 can be
changed by tilting an arbitrary side of a rotatable stage (not
illustrated) or an arbitrary side of a target stage (not
illustrated) on which the substrate to be inspected 24 is
placed.
[0032] The control unit 21 comprises the following: a total control
unit 85; a storage means 81 for storing a signal from the secondary
electron detector part 62; an image processing circuit 82 for
processing the stored digital signal; an inspection-condition
setting unit 83 for setting processing parameters of the image
processing circuit 82; and a defect data buffer 84 for storing
defect information that is obtained as a result of the processing
of the image processing circuit 82.
[0033] The operation unit 20 is provided with a display screen for
displaying an image obtained, and for displaying tools for various
instructions relating to inspection. The operation unit 20 also
includes the following: a first stored-image display button 53 and
a second stored-image display button 54, each of which is a tool
for causing the operation unit 20 to display an image stored in the
storage means 81. A comparison-result display button 55 causes the
operation unit 20 to display the result of image comparison to
extract a defect; and defect display button 56 causes the operation
unit 20 to display a defect image. According to these instructions,
a selected electron-beam image, a defect image, and the like, are
displayed in an image display window 88 on the display screen.
Further, the operation unit 20 displays a map window 87 used when
giving an instruction to display a position of an inspection area
on the substrate to be inspected 24, and an instruction to set an
inspection area. Below the map window 87 and the image display
window 88, the following are provided: an image-obtaining
instruction button 89 used to give an image obtaining instruction;
an image-processing instruction button 90 used when adjusting image
processing, for example, the contrast of an obtained image; and a
processing-condition setting button 91 used to set various kinds of
conditions such as electronic optical conditions, stage control
conditions, and vacuum pumping conditions. In addition, although it
is not illustrated, a mode switching unit for switching a mode
according to contents of operation including an inspection mode and
a recipe mode may also be provided.
[0034] In this specification, although these buttons illustrated in
FIG. 2 and the other buttons illustrated in FIG. 4, FIG. 5, and
FIG. 6 are explained as Graphical User Interface (GUI) on the
display, for example, it is possible to provide a plurality of
switches on the control unit 21 instead of the GUI.
[0035] Operation commands, and operating conditions, of each part
of the circuit-pattern inspection apparatus 16 are inputted and
output from the operation unit 20. Various conditions are inputted
beforehand into the operation unit 20 so that these conditions can
be set according to an object. Here, the conditions include an
acceleration voltage at the time of the generation of an electron
beam; the deflection width of the electron beam; the deflection
speed; the signal capturing timing of the secondary electron
detector part; and the moving speed of the stage.
[0036] In a manufacturing process for forming semiconductor devices
in a semiconductor wafer, finding a defect and studying its cause
early, and then taking early measures against the cause, will lead
to an improvement in production yield. Therefore, at the time of
the automatic inspection to detect a defect of a circuit pattern on
the semiconductor wafer, high inspection speed is indispensable for
circuit-pattern inspection apparatuses. As for the inspection
apparatus using an electron beam as described in this embodiment,
scanning at a low speed an electron beam having an electron-beam
electric current of the order of pA, or scanning such an electron
beam many times, or superimposing images on each other, as is
performed with the usual SEM, results in very low inspection speed.
Moreover, with the object of avoiding electrification of an
insulating material, the number of times an electron beam is
scanned should be once, or several times, at a high speed.
[0037] The circuit-pattern inspection apparatus in this embodiment,
therefore, is so devised that an image is formed by scanning only
once an electron beam having for example an electric current of 100
nA which is about one hundred times larger than that of the usual
SEM. This is based on the assumptions that the scan width is 100
.mu.m and one pixel is equivalent to 0.1 .mu.m.sup.2, and that a
period of time during which scanning is performed once is 1
.mu.s.
[0038] Incidentally, a thermal field emission electron source of a
diffusion supply type is used as the electron gun 25. Because using
this electron gun 25 makes it possible to ensure a stable
electron-beam electric current as compared with the conventional
tungsten (W) filament electron source and the cold field emission
electron source, an electron-beam image with little fluctuations in
brightness can be obtained. Additionally, because this electron gun
25 allows an electron-beam electric current to be set at a large
value, high-speed inspection can be achieved.
[0039] A retarding power supply 73 applies a negative voltage to
the substrate to be inspected 24. Depending on the structure, and a
material, of a circuit pattern formed on the substrate to be
inspected, there is a case where it is desirable to change the
energy of the primary electron beam 34. By adjusting the voltage of
the retarding power supply 73, it is possible to reduce the speed
of the primary electron beam 34, and thereby to adjust the
electron-beam irradiation energy applied to the substrate to be
inspected 24 to an optimum value without changing the potential of
the electron gun 25. The secondary electron 71, which has been
emitted by irradiating the substrate to be inspected 24 with the
primary electron beam 34, is accelerated by a negative voltage
applied to the substrate to be inspected 24.
[0040] An ExB deflector 33 is placed above the substrate to be
inspected 24. The secondary electron 71 accelerated by this ExB
deflector 33 is deflected in a given direction. The amount of the
deflection can be adjusted by the voltage applied to the ExB
deflector 33 and the intensity of a magnetic field. In addition,
this electromagnetic field can be changed in conjunction with the
negative retarding voltage applied to the substrate to be inspected
24. The secondary electron 71 deflected by the ExB deflector 33
collides with the reflector plate 32 under given conditions. This
reflector plate 32 has such a shape that a shield pipe and a cone
are united with each other in the scanning deflector 30 used for
the primary electron beam 34 with which the substrate to be
inspected 24 is irradiated. When the accelerated secondary electron
71 collides with this reflector plate 32, a second secondary
electron 72 having the energy ranging from several eV to 50 eV is
generated from the reflector plate 32.
[0041] As described above, the secondary electron 71, which has
been generated while the substrate to be inspected 24 is irradiated
with the primary electron beam 34, is accelerated and then collides
with the reflector plate 32, with the result that the second
secondary electron 72 is generated. The secondary electron detector
35 is devised to detect the second secondary electron 72 in
conjunction with the timing of scanning the primary electron beam
34.
[0042] The secondary electron detector 35, the preamplifier 36, the
A/D converter 37, a preamplifier driving power supply 42, an A/D
converter driving power supply 43, and a reverse bias power supply
44, are floating in a positive potential by a high voltage power
supply 41. The second secondary electron 72, which has been
generated as a result of the collision with the reflector plate 32,
is introduced into the secondary electron detector 35 by an
absorption electric field generated by a positive potential.
[0043] In the secondary electron detector part 62, a detection
signal coming from the secondary electron detector 35 placed in the
vacuum-pumped inspection cabinet 17 is transmitted to the
preamplifier 36 placed outside the inspection cabinet 17, and then
the detection signal is amplified. The A/D converter 37 converts
the amplified detection signal into digital data. The A/D converter
37 is devised to immediately convert the analog signal, which has
been detected by the secondary electron detector 35 and then has
been amplified by the preamplifier 36, into a digital signal, and
then to transmit the digital signal to the control unit 21. The
detected analog signal is first digitized immediately after the
detection, and then the digital signal is transmitted. Accordingly,
it is possible to obtain a high-speed signal with a high SN
ratio.
[0044] When performing inspection, the X stage 47 and the Y stage
46 are made stand still, and then the primary electron beam 34 is
two dimensionally scanned on the substrate to be inspected 24. In
another way, while the X stage 47 or the Y stage 46 is continuously
moved at a constant speed, the primary electron beam 34 is scanned
in an approximately right-angle direction. If a specific,
relatively small area is inspected, the former inspection method in
which inspection is performed with the stages being made stand
still is effective; and if a relatively wide area is inspected, the
inspection method in which inspection is performed while the stage
is continuously moved at a constant speed is effective.
[0045] In order not to irradiate the substrate to be inspected 24
with the primary electron beam 34, instead of the scanning
deflector 30, the blanking deflector 28 is used for the deflection,
and the iris 29 is used for the shielding. This enables not only
alternating scanning but also one-way scanning when scanning the
primary electron beam 34, which makes it possible to control the
degree to which the substrate to be inspected 24 is irradiated with
an electron beam.
[0046] With the object of knowing where the primary electron beam
34 is applied, it is important to perform position measurement when
the X stage 47 and the Y stage 46 move. In this embodiment, a
length measuring machine by means of laser interferometry is used
as the position-monitor length measuring machine 48. Positions of
the X stage 47 and of the Y stage 46 can be monitored in real time,
and its information is transmitted to the control unit 21.
Moreover, other than the position-monitor length measuring machine
48, the circuit-pattern inspection apparatus 16 is so devised that
data including the number of revolutions of a motor for driving the
X stage 47 and the Y stage 46 is also measured and then the data is
transmitted from a driver of the motor to the control unit 21. The
control unit 21 can accurately keep track of an area where the
primary electron beam 34 is applied, and its position according to
the data. If there is a discrepancy in position at which the
primary electron beam 34 is applied, position correction data is
transmitted from the control unit 21 to the correction control
circuit 58, and a corrected deflection signal is transmitted to the
scanning signal generator 59. This makes it possible to correct the
discrepancy in position in real time.
[0047] As the inspected-substrate height measuring machine 49, an
optical measuring instrument which adopts a measuring method other
than that using an electron beam is used; for example, a
laser-interferometry measuring instrument, or a reflected-light
type measuring instrument which measures a change by a position of
reflected light, is used. The inspected-substrate height measuring
machine 49 is devised to measure a height of the substrate to be
inspected 24 placed on the X stage 47 and the Y stage 46 in real
time. In this embodiment, the following method is used: irradiating
the substrate to be inspected 24 with narrow white light, which has
passed through a slit, over a transparent window; detecting a
position of its reflected light by a position detection monitor;
and calculating the amount of change in height from fluctuations in
position.
[0048] On the basis of the measured data of this
inspected-substrate height measuring machine 49, the focal length
of the objective lens 31 used to thinly narrow the primary electron
beam 34 is dynamically corrected, which makes it possible to emit
the primary electron beam 34 that is always focused on a
noninspection area. In addition, a warp, the height, and the
distortion, of the substrate to be inspected 24 are measured in
advance before irradiating with an electron beam. Therefore, on the
basis of such data, it is also possible to set correction
conditions for each inspection area of the objective lens 31.
[0049] An image signal of the substrate to be inspected 24, which
has been detected by the secondary electron detector 35, is
amplified by the preamplifier 36, and is then digitized by the A/D
converter 37. After that, the digital image signal is converted
into a light signal by an optical conversion means 38, and is then
transmitted by an optical transmission means 39. The light signal
is converted into an electric signal again by an electric
conversion means 40 before the electric signal is stored in a
storage means 81 of the control unit 21.
[0050] The defect detection is performed by comparing the first
image signal stored in the storage means 81 with the second image
signal to extract the difference between both of the image signals.
The image processing circuit 82 performs various kinds of image
processing with the object of aligning the first image signal with
the second image signal, standardizing a signal level, and
eliminating a noise signal. Then, comparison operation of both
image signals is performed. An absolute value of a difference image
signal resulting from the comparison operation is compared with a
given threshold value on a pixel basis. If a signal level of the
difference image signal is larger than the given threshold value,
this pixel is judged to represent a defect candidate, and
accordingly the image signal, position data, and the like, are
stored in the defect data buffer 84. The image processing circuit
82 also can calculate a correlative value between the first image
signal and the second image signal.
[0051] Next, condition settings at the time of inspection will be
described. On a screen of the operation unit 20, the map window 87
for displaying a position of a current stage and the image display
window 88 for displaying an image of the optical microscope 19 are
displayed. Clicking on this map window 87 will move the X stage 47
and the Y stage 46 so that a position at which conditions are set
can be selected. In addition, clicking on the image-obtaining
instruction button 89 of the operation unit 20 causes the secondary
electron detector part 62 to detect the secondary electron 71 or
the second secondary electron 72, which is generated by irradiating
the substrate to be inspected 24 with the primary electron beam 34.
After that, the detected electron is converted into a digital
signal, and a digital image of a given area is then stored in the
storage means 81.
[0052] By clicking on the processing-condition setting button 91 of
the operation unit 20, image processing conditions are set. By
clicking on the image-processing instruction button 90, image
processing is executed under the set processing conditions.
Further, on the basis of the conditions set by the
inspection-condition setting unit 83 of the control unit 21, the
digital image stored in the storage means 81 is processed by the
image processing circuit 82 to extract a defect, and coordinate
data of the defect and an image of the defect are stored in the
defect data buffer 84.
[0053] FIG. 2 is a diagram illustrating a screen configuration of
the operation unit 20. FIG. 2 illustrates the operation unit 20
shown in FIG. 1 in detail. In the map window 87 of FIG. 2, an area
where an image has been obtained to set inspection conditions is
displayed in a manner that this area can be distinguished from the
others. Clicking on the area and then pressing a zoom in button 101
will enlarge the area. By clicking on a symbolized defect in the
enlarged area displayed in the map window 87, an image of the area
including the defect is displayed in the image display window
88.
[0054] The above-mentioned work is repeated to search for
inspection conditions that are suitable for the inspection. When
the condition check at one position is completed, the map window 87
is displayed on a reduced scale again by pressing a zoom out button
102. Next, the image display window 88 is switched to an optical
microscope image, and a condition-setting position is selected
again. Then, the steps from the obtaining of an image to the
condition settings are repeated.
[0055] Next, various kinds of parameters required to execute an
inspection will be described. Such parameters include parameters
that are specific to a substrate to be inspected, and parameters
that determine operating conditions of an apparatus.
[0056] The parameters which are specific to a substrate to be
inspected can be roughly grouped into two kinds. One is a parameter
called a "classification file", which is not changed by a layer in
the middle of a manufacturing process. This kind of parameter
includes, for example, the size of the substrate to be inspected
24; a shape of an orientation flat or of a notch; the exposure shot
size of a semiconductor product; the chip (or die) size; a memory
cell area; and the iteration unit size of the memory cell. These
parameters are stored in a table as a "classification file".
[0057] The other is a parameter called a "process file", which
needs adjustment because a state of a surface material and a state
of a surface shape differ depending on a layer in the middle of the
manufacture process. This kind of parameter includes, for example,
electron-beam irradiation conditions; various kinds of gains in the
detection system; and image-processing conditions for detecting a
defect. These parameters are written as the "process file".
[0058] At the time of inspection, by specifying the "classification
file" and the "process file", it is possible to call inspection
conditions corresponding to a specific semiconductor product and a
specific manufacturing process. In this embodiment, the
"classification file" and the "process file" are designated as a
"recipe" as a whole. In addition, a series of operations relating
to inputting, and writing, of these various kinds of parameters is
called "recipe creation".
[0059] A recipe creation method, and an operation screen used to
execute the recipe creation method, will be described as below. The
recipe creation screen shown in FIG. 2 is roughly divided into five
areas. An area 103 is located in the upper part of the screen.
Information displayed in the area 103 includes an apparatus name,
and an apparatus ID, and a classification file name and a process
file name, both of which are used as a recipe name. In an area 104,
guidance for describing the operation and the states is displayed.
In an area 105 located in the center of the screen, displayed
contents change in response to the operation and a progress state.
In the example of FIG. 2, the map window 87 and the image display
window 88 are displayed. In an area 106 located on the right side
of the screen, operation buttons which are required for a plurality
of screens in common are displayed. As examples of the operation
buttons, there are "Start", "Load wafer", "Unload wafer", "Print
screen", "Save file", "Save image", "End", and the like. For
example, pressing the save file button will display a screen that
is used to specify names of a classification file and of a process
file into which a recipe currently being created is saved. In
addition, pressing the save image button will display a screen used
to specify a name that is required to save as an image file an
image currently being displayed.
[0060] A mode name is displayed in an operation area 107 in the
lower part of the screen. For example, pressing the inspection
button 108 will enter an inspection mode for executing automatic
inspection; and pressing the create recipe button 109 will enter a
recipe creation mode for setting inspection conditions such as
inputting of the above-mentioned parameters.
[0061] Incidentally, even in the case of another apparatus, a model
of which is the same as that of the apparatus currently being used,
creating a recipe once suffices so long as the same recipe can be
used, which makes the apparatus convenient to use. Usually, a
recipe cannot be transmitted to another apparatus. The inspection
result, however, can be inputted into any apparatus. A defect
information file which is the inspection result includes inspection
conditions. Therefore, inputting the defect information file into
another apparatus eliminates the need for newly creating inspection
conditions from the beginning, which makes the apparatus convenient
to use. If in the inspection conditions there is a condition that
should be changed, it is possible to change the condition in the
recipe creation mode described below.
[0062] FIG. 3 is a flowchart illustrating processing of the recipe
creation mode. Additionally, FIG. 4 is a diagram illustrating a
screen used in the recipe creation mode. In the screen shown in
FIG. 2, when pressing the create recipe button 109, the screen of
the operation unit 20 shown in FIG. 1 is switched to a screen shown
in FIG. 4, which is displayed at the time of recipe creation. A
start button 401 is pressed in this screen (step 301 in FIG. 3).
Next, a shelf number of a cassette in which a substrate to be
inspected is stored (the shelf number is displayed in a
shelf-number display area 402) is specified to call a recipe file
(step 302). Then, in a classification condition entry area 403,
whether classification conditions are newly created or modified is
specified; in a lot ID input area 404, a lot ID is inputted; and in
a wafer ID input area 405, an ID of the substrate to be inspected
is inputted (step 303). When changing a classification condition,
regardless of whether or not the substrate to be inspected is
loaded into the apparatus, it is possible to change a recipe
creation condition. The change is usually made with the substrate
to be inspected being loaded.
[0063] Next, by pressing the load wafer button 406, the substrate
to be inspected is loaded from the cassette to the inspection
apparatus (step 304). The inspection apparatus detects the
orientation flat, or the notch, of the substrate to be inspected.
After that, the inspection apparatus uses a target-wafer holder to
hold the substrate to be inspected in a target-wafer exchanging
chamber, and then transfers the target-wafer holder to a stage of
an inspection cabinet. Further, on the screen shown in FIG. 4, a
message 408 indicating that a wafer is being loaded is displayed
with a color different from the others.
[0064] Next, the inspection apparatus performs calibration of a
beam (step 305). Calibration conditions require calibration on the
basis of default recipe-file conditions stored beforehand.
Accordingly, the inspection apparatus performs correction required
for this substrate to be inspected, and then stores the changed
conditions. The calibration conditions include irradiation
conditions of the primary electron beam; the amount of correction
of deflection; the amount of correction of reference coordinates;
and the amount of correction of a focus parameter.
[0065] Next, the image contrast is checked by irradiating a
specified position of the substrate to be inspected with the
primary electron beam, and then a focus and astigmatism are
readjusted (step 306). As shown in FIG. 2, an image is displayed in
the image display window 88 of the area 105 in the center of the
screen. If the sufficient contrast cannot be obtained (step 307),
the process returns to the previous step to change the
electron-beam irradiation conditions (step 305). Here, the
specified irradiation conditions, and the focus and astigmatism
conditions, are stored in a process file as recipe parameters.
[0066] After the irradiation conditions of the primary electron
beam have been determined and the contrast has been checked,
die/shot matrix are checked (step 308). To begin with, the size,
and a matrix (array), of a shot of the substrate to be inspected,
and the size, and a matrix (array), of a die (chip) in the shot are
inputted. Next, whether or not there is a shot or a die around the
perimeter of the substrate to be inspected is specified. The
matrices of the shot and of the die are stored as parameters in the
recipe file.
[0067] Next, alignment conditions are inputted, and thereby the
alignment is executed (step 309). In more detail, the alignment
includes the following steps: (1) specifying a plurality of
alignment chips; (2) moving to the starting point of a first chip;
(3) switching to an optical microscope monitor; (4) manually moving
to an alignment mark position of the first chip; (5) writing an
optical image; (6) switching to a SEM image mode; (7) manually
making a fine adjustment to an alignment mark position; (8) writing
a SEM image; and (9) writing alignment coordinates.
[0068] In more detail, the execution of the alignment includes the
following: (1) move of a first point; (2) input of an image,
search, and matching; (3) move of a second point; (4) input of an
image, search, and matching; (5) move to a remaining point, search,
and matching; and (6) correction of inclination, correction of a
position, and correction of a chip interval.
[0069] Further, in more detail, the offset settings of the starting
point of the chip include the following: (1) moving to a last point
alignment mark; (2) specifying an alignment mark position (SEM
image mode); (3) moving to the starting point of a first point
chip; (4) specifying a position of the starting point of the chip
(SEM image mode); and (5) calculating and writing an offset between
the starting point and the alignment mark. The offset of the
starting point of the chip is the distance from the alignment
coordinates to starting point coordinates of the chip including the
mark.
[0070] Thus, a value of the offset from specified pattern
coordinates for alignment to the starting point of the chip is
inputted, and this offset value is then written into the process
file as an alignment parameter. As for the recipe creation, there
are many parameters for specifying coordinates used to execute
various kinds of processing on a wafer. Accordingly, the alignment
conditions are first determined and written, and then processing up
to the alignment is executed.
[0071] Next, a memory cell area in the chip is set (step 310). In
more detail, the memory cell area settings include the following:
(1) inputting a cell area; (2) inputting the cell pitch; and (3)
memorizing (1) and (2) to the storage means 81. The cell area is
inputted by use of an optical-microscope image and an electron-beam
image.
[0072] Next, a die area is set (step 311). In more detail, the die
area settings include the following: (1) inputting a die area; (2)
inputting a die noninspection area; and (3) memorizing (1) and (2)
to the storage means 81. The die area is also inputted by use of an
optical-microscope image and an electron-beam image.
[0073] Next, an inspection area is specified (step 312). When
specifying the inspection area, two kinds of inspection areas can
be specified, one is an inspection die and another is an inspection
area in the die. If it is not necessary to inspect all dies, or if
only a specific area in a die needs to be inspected, it is possible
to specify it arbitrarily. Moreover, with the object of further
specifying, in the specified area, an inspection area or an area
that needs not be inspected, it is also possible to specify an
inspection sampling rate and a skipping method. Additionally, it is
also possible to specify inspection in the X direction, or
inspection in the Y direction. Data as to the die area and the
inspection area are stored as parameters in the process file.
[0074] As described later, the operation unit 20 is programmed to
allow to arbitrary selection of any dies on the semiconductor wafer
as the die to be inspected and the reference die. The operation
unit 20 is programmed to permit selection of a die which is not
adjacent to any side of a die to be inspected as a reference die.
For one example, the any side of die is lateral side and/or
longitudinal side of a rectangle die. The die which is not adjacent
to any side of die to be inspected is separated from the die to be
inspected by at least one intervening die. In following
explanation, although the operation unit 20 is explained as it
selects the die which is not adjacent to any side of the die to be
inspected as a reference die. Of course, at other times the
operation unit 20 can also select the adjacent die as the reference
die.
[0075] Also, the operation unit 20 is programmed to permit setting
of an order of dies scanned in sequence so that a die be inspected
and a separate reference die are consecutive in the scanning
sequence. Furthermore, the operation unit 20 is programmed to
permit selection of a die to be inspected and a reference die
separated from each other on a semiconductor wafer with at least
one die which is intervening in one or two directions of straight
outlines of the dies, between the die to be inspected and the
reference die.
[0076] The processing unit 20 has a computer or other processor
which controls the circuit-pattern inspection apparatus built-in.
The computer performs a sequence of operations according to
executable code embodied in a readable medium when the program
installed in the processing unit 20. And the sequence is performed
by setting the die to be inspected, the reference die, and the
order of dies scanned in consecutive sequence.
[0077] Aspects of the methods outlined above may be embodied in
software, e.g. in the form of program code executable by the
computer or other programmable device. Such software typically is
carried on or otherwise embodied in a medium or media. Terms such
as "readable medium" used herein refer to any medium that
participates in providing instruction and/or data to a programmable
processor, such as a CPU for execution or other processing.
[0078] On the completion of the specification of an inspection
area, the process proceeds to calibration settings for adjusting
brightness at the time of the inspection (step 313). The
calibration is performed as follows: first obtaining an image, and
on the basis of the distribution of its brightness, executing gain
adjustment, and brightness correction, of hardware in response to
the quantity of signal. In actuality, the calibration is performed
by specifying a die where the calibration is performed, and
coordinates in the die. The coordinate values, the gain of
brightness, and the offset value, which are used for the
calibration, are stored as parameters in the process file.
[0079] Next, under the various kinds of conditions that have been
set up to this time, an image is actually obtained, and thereby
image processing conditions for detecting a defect are set (step
314). In the first place, when obtaining the image, a kind of a
filter which is used for a detection signal is selected. Then,
under the same conditions as those of the inspection, an image of a
small area in one chip is actually obtained. Here, on the
assumption that for example a scan width of a primary electron beam
is 100 .mu.m, the small area indicates an area having a length
equivalent to one chip with respect to its width. After obtaining
an image, a threshold value used to judge the image to be a defect
is inputted, and thereby an image at a position, which has been
judged to be a defect, is displayed. Optimum inspection conditions
are determined by repeating this. This series of work is called
"small-area trial inspection". The parameters including the
threshold value and a file, which have been set here, are stored as
parameters in the process file. The various kinds of inputs as
described above allow the various kinds of parameters required for
the inspection to be set.
[0080] However, as for a semiconductor wafer that is the substrate
to be inspected, there are the unevenness inside a wafer surface
and the unevenness caused by the process among production lots.
Therefore, the settings of the image processing conditions only for
the small-area trial inspection are not sufficient, and accordingly
it is necessary to determine the threshold value for the defect
judgment taking these unevenness factors into consideration. For
this reason, after evaluating the result of the inspection in which
the created recipe has been used, the various parameters are
written into the classification file and the process file.
[0081] To begin with, the final inspection is performed by use of
the created recipe file (step 315). In more detail, the final
inspection includes the following steps: (1) continuously moving a
stage at a constant speed while monitoring its position and the
height; (2) scanning a beam with the correcting in real time
(followed by the correction of a stage and a Z sensor); (3)
detecting a secondary electron, A/D conversion, and inputting an
image into a memory; (4) image processing, and comparison for
judgment; (5) correcting a beam on a N stripe basis; and (6)
displaying the number of the defects and defect positions.
[0082] Next, the result is reviewed to check a defect detection
level and an improper detection level (step 316). If the conditions
are proper, the various kinds of parameters of the recipe, which
have been inputted up to this time, are written into the
classification file displayed in the classification-file display
area 409 and the process file displayed in the process-file display
area 410 in FIG. 4 (step 317). Then, by pressing the unload wafer
button 407, the substrate to be inspected is unloaded (step 318).
After that, the end button 411 is pressed to end the recipe
creation mode (step 319).
[0083] Here, points which have been improved by the recipe creation
mode will be described. As far as the conventional inspection
apparatuses are concerned, an operator specifies a die to be
inspected only as an area; to be more specific, what can be changed
are only directions (for example, X direction, and Y direction) of
comparison. The order of comparison checking inside of the
specified die cannot be changed at all.
[0084] For example, even if comparing between dies at edges in the
X direction, or comparing between edges in the Y direction, in a
piece of semiconductor wafer, makes it possible to know the
unevenness in the semiconductor wafer, the conventional inspection
apparatus can compare only adjacent dies. This means that the
conventional inspection apparatuses cannot know the unevenness in
the semiconductor wafer. Moreover, when manufacturing two or more
kinds of semiconductor chips in a piece of semiconductor wafer,
comparison checking between different kinds of semiconductor chips
is meaningless. Nevertheless, the conventional inspection apparatus
performs the comparison checking of only adjacent semiconductor
chips in sequence. Therefore, the obtained inspection result will
contain useless defect information (false report) that is expected
to be wrong in advance and should be eliminated from the inspection
result.
[0085] For this reason, in this embodiment, an operation screen is
provided so that when the operator sets an inspection area, the
operator can arbitrarily specify the order of dies to be inspected.
As a result, the comparison checking of dies is executed in the
specified order.
[0086] FIG. 5 is a diagram illustrating a screen for setting the
inspection order of dies. In the step 312 of the flowchart shown in
FIG. 3, the operator uses the screen shown in FIG. 2 to specify the
order of dies to be inspected. In an inspection-area specifying
area 110 of the screen shown in FIG. 2, "random" is specified.
Then, dies 501 displayed in the map window 87 are clicked in turn
as shown in FIG. 5. The specified dies are displayed as their
numbers in of the selected order. Moreover, if arrows as shown in
FIG. 5 are displayed to indicate the specified order, the order of
inspection can be easily known. The inspection is performed by
comparing between dies according to this order. In addition, if
there is a die to be inserted after the order has been set,
insertion processing is allowed. In this case, an insert button 502
shown in a FIG. 5 is first pressed, and a die immediately before
the insertion is clicked, and then a die to be inserted is clicked.
The die numbers after the inserted die are renumbered in a manner
that the die numbers are all incremented by one.
[0087] Moreover, if the operator makes a mistake during or after
the specification of dies, or if there is a die that need not be
inspected, pressing a delete button 503, and then clicking a die to
be deleted, will perform deletion processing. As a result, the die
numbers after the deleted die are automatically decremented by
one.
[0088] Thus, if there is a difference in distribution in the
semiconductor wafer caused by production (for example, in a case
where the inspection result differs when comparing between dies in
the circumferential part), it is possible to perform such
inspection that only dies in the circumferential part is set and
compared at a time. Therefore, it is possible to check whether or
not there is a difference in distribution in the semiconductor
wafer, and what is more, a die area other than the circumferential
part, which does not require inspection, is not inspected.
Accordingly, the amount of time is shortened, enabling efficient
inspection.
[0089] Additionally, in order to prevent the inspection result from
including a false report, not specifying dies which should not be
compared can avoid misjudgment that recognizes the inspection
result as an abnormal condition although it is a false report.
[0090] According to one example, the image data of the die to be
inspected and the reference die are proceeded to obtain a
correlative value with the image processing circuit 82. The control
unit 21 classifies the image data based on whether the correlative
value is under a predetermined value or not, and the classified
data is memorized in the storage means 81. The image data
classified as being under the predetermined value indicate a
possibility that there is a defect changing gradually ranging from
the die to be inspected to reference die. Therefore it is possible
to determine whether there is the defect changing gradually between
dies based on the classified data.
[0091] And according to the other example, in case of the
correlative value being low nevertheless it can be determined that
there is no specific part defect (ex. a snapping of line pattern,
or foreign body) by calculating the difference image signal, there
may be a defect changing gradually between dies. Because the
correlative value between the images may change by an existence of
the specific part defect, there is every possibility that the
specific part defect causes a declination of the correlative value.
However, in case of the correlative value being low nevertheless it
can be determined that there is no specific part defect, it is
considered that a difference factor contributes to fall the
correlative value. Therefore, it can confirm whether there is the
defect changing gradually between dies by judging based on the
correlative value between the dies after confirming there is no
specific part defect.
[0092] Furthermore, the comparing between separate dies can make it
possible to confirm a condition of whole semiconductor wafer as a
deformation of the semiconductor wafer or a condition of a resist
on the semiconductor device. For example, when dies of four corners
on the semiconductor wafer are compared in order of left upper,
right upper, right lower, left lower, and left upper of the dies
from the left upper die as starting point, there may be the
deformation in a gap between left and right on the semiconductor
device, in case of correlative values between left upper and right
upper dies, and between left lower and right lower dies being under
pre-determined value. The comparing between separate dies can make
it possible to do the confirmation of whole semiconductor wafer
which can not be accomplished by comparing between adjacent
dies.
[0093] As discussed above, the inspection result can be obtained
selectively by selecting a die for the reference image that is not
adjacent to the die to be inspected. As a result, an efficiency of
inspection can be improved.
[0094] As another example, it is possible to switch techniques of
comparison inspection as a need by comparing between separate dies.
For example, if a correlative value obtained by comparing between a
die on one end of a semiconductor wafer and a die on the other end
of the semiconductor wafer is higher than a predetermined value, it
can decide that there is no gentle change extending from the one
end to the other end. In this case, for example, it is possible to
improve the efficiency of inspection by reducing a number of the
dies to be inspected. As an example, when the correlative value
between the one end and the other end is more than the
predetermined value, it may be able to inspect a plurality of dies
surrounding the one end of die or the other end of die selectively.
When the correlative value is under the predetermined value, it may
be appropriate conduct a detailed inspection of all dies on the
semiconductor wafer.
[0095] As above mentioned, the techniques of comparison inspection
are switched by the setting of a die as the reference die that is
not adjacent to the die to be inspected and the determining whether
the correlative value between the reference die and the die to be
inspected is under the pre-determined value or not. In this case,
it is possible to improve an efficiency of the inspection when the
correlative value between the reference die and the die to be
inspected is high, and it is possible to maintain an accuracy of
inspection when the correlative value is low.
[0096] It is also possible to switch from comparing between
adjacent dies to comparing between separate dies, when a designed
tendency is found in correlative values between the dies is found
during comparing between adjacent dies.
[0097] For example, the setting of the switching of the comparison
techniques can be performed by a setting a change from the "ALL
OVER IN THE X DIRECTION" to "FIX REFERENCE DIE" and a setting the
correlative value on the display as illustrated in FIG. 5 or FIG.
6.
[0098] Next, screen operation used when fixing dies to be compared
will be described. As described above, as far as the conventional
inspection apparatuses are concerned, an operator specifies a die
to be inspected only as an area, and accordingly the order of
comparison checking inside the specified die cannot be changed at
all. In this case, a reference die with which the die to be
inspected is compared is an adjacent die, or a die that depends on
the order of the stage movement. If there is the moderate
distribution of states of produced dies in a semiconductor wafer
caused by the process, this embodiment is effective. To be more
specific, if a comparison is made between adjacent dies, a slight
change is detected, and this change is not recognized as an
abnormal condition. However, if it is necessary to detect such an
abnormal condition that can be found out by comparing between edges
of the semiconductor wafer, or by comparing the circumference with
an area around the center, or by comparing the beginning of the
inspection with the end of the inspection, this embodiment is
effective.
[0099] FIG. 6 is a diagram illustrating a screen used for fixed
settings of a reference die. In the inspection-area specifying area
110, "Fix reference die" is selected; and in the map window 87
shown in FIG. 2, a reference die 601 which is referred to at the
time of comparison is specified. At the time of inspection, the
inspection is always performed by comparing with this specified
reference die 601.
[0100] Besides the reference die 601, a plurality of reference dies
can also be specified (reference die 602). For example, as shown in
FIG. 6, the reference die 601 is compared with an inspection die
row 603; and the reference die 602 is compared with an inspection
die row 604.
[0101] In this manner, if there is the moderate distribution of
abnormal conditions in a semiconductor wafer caused by the
production, performing inspection with the reference die being
fixed makes it possible to detect the moderate distribution of
abnormal conditions in the whole semiconductor wafer. Further,
because the die in the semiconductor wafer is always compared with
the fixed die, it becomes possible to detect the absolute
distribution instead of the relative distribution of abnormal
conditions of the die.
[0102] Incidentally, devising a method for selecting a fixed die
makes it possible to adjust the defect detectivity. To be more
specific, by using a die containing many defects as a reference die
to be compared, it is possible to reduce as an offset the total
number of defects in the semiconductor wafer. As a result, an
uncritical defect is not detected, which reduces the number of
target defects to be reviewed after the inspection, making it
possible to improve the efficiency in review.
[0103] Moreover, even if two or more kinds of semiconductor chips
are provided on one semiconductor wafer, it is possible to inspect
only required chips, which enables efficient inspection.
[0104] Such settings of targets to be inspected can be applied not
only to the SEM image in this embodiment but also to an image of an
optical microscope and an image obtained with a laser beam in the
same way of thinking.
[0105] As described above, because the operator can quickly perform
the inspection of a chip, the sampling inspection of a wafer, and
the like, while viewing the screen, it is possible to quickly
detect a defect across the whole product or a defect in a specific
area. Furthermore, because fluctuations of process conditions are
reliably detected, the fluctuation can be fed back to the process
and also to the adjustment of the difference in man-hour and of an
expenditure budget.
[0106] If the above-discussed inspection apparatus is applied to a
substrate product process, it possible to detect an abnormal
condition relating to a product, an apparatus, or conditions at an
early stage and with a high degree of accuracy by referring to a
screen. Accordingly, it is possible to quickly take measures
against the abnormal condition during the substrate manufacture
process. As a result, the defective rate of a semiconductor device
and of other substrates can be reduced, enabling an improvement in
productivity.
[0107] While the invention has been described in its preferred
embodiments, it is to be understood that the words which have been
used are words of description rather than limitation and that
changes within the purview of the appended claims may be made
without departing from the true scope and spirit of the invention
in its broader aspects.
* * * * *